KNOWLEDGE IS POWER.
Power concentrates wherever knowledge becomes scarce, private, or illegible. Education, source code, research, and the means to think must remain radically accessible.
↗INDEPENDENT SIGNAL / LIVE
PERSONAL FREQUENCY / 00
LEONARDO MARCIANO / CTO / AI ENGINEER / HACKER
Marciano Theory is where I publish the ideas I use to understand intelligence, freedom, longevity, power, and the systems quietly deciding our future.
MY IDEAS. MY CONTRADICTIONS. MY PUBLIC RECORD.
FOLLOW THE ARGUMENT
01 / WHO IS TRANSMITTING
I am Leonardo Marciano—an artificial-intelligence engineer, hacker, liberal, longevity researcher, and stubborn believer that people deserve access to the machinery shaping their lives.
I founded Livvay to pursue the most literal form of human agency: more healthy time. I study longevity and immortality because death is a limit worth interrogating, not romanticizing.
In AI, I ask the same political question: who owns the intelligence? Knowledge is power. Models should be open, infrastructure decentralized, and people less dependent on institutional permission.
This site is not a company façade or a portfolio. It is a living notebook—part manifesto, part bibliography, part open channel.
PUBLIC REACH / TOOLZZ
As CTO of Toolzz, I help shape education and artificial-intelligence infrastructure used across companies, institutions, and everyday digital experiences.
Reach figures supplied by Leonardo Marciano. Supporting methodology will be added to the final public record.
THINK IN PUBLIC.
KEEP THE BLACK BOX OPEN.
02 / CURRENT THESES
Six positions connecting liberalism, open intelligence, decentralization, hacker culture, longevity, and the distribution of power.
Power concentrates wherever knowledge becomes scarce, private, or illegible. Education, source code, research, and the means to think must remain radically accessible.
↗AI too important to inspect, modify, or run without permission is power rented from a gatekeeper. Models, research, and infrastructure should move toward openness.
↗No single company, government, or cloud should own the effective limits of machine intelligence. Distributed compute and plural models make capture harder.
↗I am a liberal because I trust individual sovereignty more than centralized permission. Technology is progress only when it expands the space in which people can choose.
↗Aging is not sacred because it is old. Biology is machinery we are still learning to read. Longevity research is the refusal to confuse current limits with natural law.
↗To open a black box is to challenge the inevitability of its design. The hacker ethic begins with curiosity and becomes political when knowledge is hoarded.
↗03 / EXECUTABLE IDEAS
Working instruments for making model decay visible, agent workflows portable, rewrites provable, and architecture enforceable.
A weekly monitor for frontier models that shows whether quality improved, held steady, or degraded against a baseline.
A local bridge that moves one coding session between the native Claude Code and Codex CLIs without replacing either interface.
A Rust parity engine for validating that observable behavior survives rewrites between languages, runtimes, and stacks.
A TypeScript architecture contract validator that enforces layers, dependency boundaries, and AST expectations in CI/CD.
SOFTWARE IS AN ARGUMENT THAT CAN BE FORKED, TESTED, AND REFUSED.
04 / MARGINALIA
Books, reports, essays, press, and public contributions. Exact credit is part of the work; unknown metadata stays visibly unknown.
THE ARCHIVE IS INCOMPLETE BY DESIGN. ACCURACY BEFORE SELF-MYTHOLOGY.
05 / MACHINE READING
Question the manifesto. Challenge a thesis. The answer is assembled only from what I have made public here.
06 / OPEN CHANNEL
Ideas, disagreements, book invitations, open AI, longevity research, editorial collaborations, and difficult questions belong here.